ζZetesis

Paving the way toward a standard for scientific due diligence in the omics era.

The verification layer between scientific claim and capital allocation.

The logic of scientific inquiry. The architecture of modern ML. The certainty of cryptographic proof.

The service

The scientific audit, engineered as infrastructure.

Zetesis is a managed due-diligence service built on proprietary ML infrastructure. Every engagement is validated by the domain experts who built the system. Every verdict ships with a zero-knowledge cryptographic attestation: verifiable by any third party and tamper-evident by construction.

Claims fail at discipline boundaries: statistically credible but clinically incoherent, biologically plausible but regulatorily untenable. Zetesis runs a pre-registered battery across each boundary.

D.01

Genetics & genomics

Causal inference, effect-size plausibility, population structure.

D.02

Statistical inference

Multiple-testing burden, model assumptions, replication.

D.03

Clinical medicine

Endpoint validity, patient-population fit, translation path.

D.04

Pharmacology

Target engagement, PK/PD, mechanism coherence.

D.05

Regulatory science

FDA/EMA precedent, trial-design feasibility, approval-path risk.

D.06

Intellectual property

Claim defensibility, freedom-to-operate, competitive encirclement.

Claim
Pre-registration
Multi-agent run
Expert validation
Attested verdict

The ML runs the investigation. The domain experts who built the system validate the verdict before release. The output is an audit-grade report delivered inside deal-timeline windows.

Applicable across: Discovery · Preclinical · Clinical · Pre-approval
An omics-platform audit in session at a verification desk
FIG.01  ·  Scrutiny
Where assertions become evidence.

Deep-tech infrastructure, purpose-built for scientific verification.

A layered architecture. Every verdict traverses each layer, in order.

L.01 Substrate

The live biomedical record

The evolving scientific corpus every layer above reasons over.

L.02 Models

Domain-adapted foundation models

Continuously trained. Accuracy compounds as the record evolves.

L.03 Orchestration

Multi-agent reasoning

Domain-specialized agents, coordinated in parallel.

L.04 Methodology

Pre-registration enforcement

The protocol locks before execution. Reproducibility by construction.

L.05 Attestation

Cryptographic proof of execution

Independently verifiable. Tamper-evident by design.

Built for the omics era of healthcare investing.

Omics-derived therapeutic platforms are now the dominant source of biotech deal flow. The claims that underpin them are themselves computational: gene–disease association scores, ML-generated target predictions, pattern-library matches in multi-omics datasets. Zetesis exists to verify these at investment grade.

A.01 · Training
Continuously trained.
The engine compounds in accuracy as the biomedical record evolves.
A.02 · Methodology
Pre-registered.
The investigation protocol locks before execution. No post-hoc reweighting.
A.03 · Proof
Cryptographically attested.
Every verdict is independently verifiable and tamper-evident by construction.

Omics is the first domain. The verification architecture extends to any scientific claim that can be audited against the live research record. Genomics today; materials, chemistry, and physics next.

The biomedical research record: the live substrate every layer of verification reasons over
FIG.02  ·  The substrate
The biomedical record, under observation.

Clinician and computational scientist, by design.

Reut Avidan, MD
Reut Avidan, MD
Co-founder · Chief Executive

Ophthalmology resident at Rambam Health Care Campus and MSc candidate in Computational Medicine at the Rappaport Faculty of Medicine, Technion. Research focus: multi-omics approaches to ocular disease, including epigenetic ocular longevity, oncology pharmacogenomics, and drug-target discovery.

Recent work: IIH pharmacogenomics pipeline (FinnGen R12, MAGMA, LDSC); medulloepithelioma DICER1 target validation (DepMap + PRISM + tumor multi-omics); presbyopia multi-omics aging atlas.
Prof. Yonatan Savir
Prof. Yonatan Savir
Co-founder · Chief Scientific Officer

Assistant Professor, Rappaport Faculty of Medicine, Technion. Information processing in biological systems; ML for drug discovery, digital pathology, biomedical signal interpretation. BSc Technion (Physics + EE); MSc/PhD Weizmann Institute; postdoc Harvard Medical School.

Methodological foundations: Taub & Savir, JCIM (2024), Smart Aggregation Framework; Daniel, Larey & Savir, IEEE EMBC (2023), DEPAS generative pathology; Savir & Tlusty, Cell (2013), ribosome as optimal decoder; Savir & Tlusty, PLOS ONE (2007), conformational proofreading.

Built with rigorous science.

Technion — Israel Institute of Technology
Ruth and Bruce Rappaport Faculty of Medicine
Rambam Health Care Campus
C.00 · Engage

Evaluating an omics-platform deal?

Zetesis is accepting charter clients from healthcare-focused venture firms, corporate venture arms, family offices, and deep-tech funds. Bring a live claim; we'll run the assessment.

Request a charter engagement